Light emitter

Abstract

Provided is a light emitter having high manufacturing stability and a high reflection suppression effect. The light emitter is monolithically integrated with the following components on one substrate: a distributed feedback type (DFB) laser having an active region formed by multiple quantum wells generating optical gain through current injection and a diffraction grating; an electric field absorption (EA) modulator having an absorption region formed of multiple quantum wells of a different composition from the DFB laser; a semiconductor amplifier (SOA) having an active region of the same composition as the DFB laser; bending the waveguide to rotate the transmission direction of the light by an angle [theta] wg; and a passive waveguide connected to the SOA and having a core having a band gap wavelength shorter than the oscillation wavelength of the DFB laser. The passive waveguide includes a tapered region and a narrow waveguide region. The tapered region is configured so as to convert the width W1 of a passive waveguide connected to the SOA into the width W2 of the narrow waveguide region, and the passive waveguide forms an angle [theta] wg with the normal to the end surface of the substrate and is in contact with the end surface of the substrate.

Description

Technical Field

[0001] The present disclosure relates to an optical transmitter, and more specifically, to a semiconductor laser element with an electric field absorption (EA) optical modulator integrated on an InP substrate, and more specifically, to an optical transmitter composed of an EA modulator, a semiconductor optical amplifier (SOA), and a distributed feedback (DFB) laser. Background Art

[0002] With the popularity of video distribution services and the increase in mobile traffic demand in recent years, network traffic has increased explosively. In the optical transmission lines that support the network, the trend is to achieve low costs through higher transmission rates, lower power consumption, and longer transmission distances. For semiconductor modulated light sources used in the network, it is also required to achieve high speed / high output while suppressing excessive power consumption.

[0003] Electric field absorption modulator integrated DFB (EADFB) lasers have higher extinction characteristics and excellent chirp characteristics than directly modulated lasers, and therefore have been used in a wide range of applications to date.

[0004] Furthermore, an EADFB laser (SOA Assisted Extended Reach EADFB Laser: AXEL) in which a semiconductor optical amplifier (SOA) is integrated has been proposed (for example, refer to Non-Patent Document 1). The AXEL can be used as an optical transmitter in optical communications.

[0005] Prior art literature

[0006] Non-patent literature

[0007] Non-patent document 1: W Kobayashi et al., "Novel approach for chirp and outputpower compensation applied to a 40-Gbit / s EADFB laser integrated with a shortSOA," Opt.Express, Vol.23, No.7, pp.9533-9542, Apr.2015 Summary of the invention

[0008] It is desirable to improve the manufacturing stability of AXEL and to achieve a sufficient reflection suppression effect at the end face of a semiconductor chip on which AXEL is formed.

[0009] The present disclosure has been made in view of such circumstances, and an object of the present disclosure is to provide a light transmitter having high manufacturing stability and high reflection suppression effect.

[0010] In order to achieve such an object, an optical transmitter of one embodiment of the present invention is an optical transmitter in which the following components are monolithically integrated on a substrate: a distributed feedback (DFB) laser having an active region and a diffraction grating formed by multiple quantum wells that generate optical gain by current injection; an electric field absorption (EA) modulator having an absorption region formed by multiple quantum wells of a composition different from that of the DFB laser; a semiconductor amplifier (SOA) having an active region of the same composition as that of the DFB laser; a bent waveguide that rotates the propagation direction of light by an angle θwg; and a passive waveguide connected to the SOA and having a core with a bandgap wavelength shorter than the oscillation wavelength of the DFB laser, wherein the passive waveguide includes a tapered region and a narrow waveguide region, the tapered region being configured to convert a width W1 of the passive waveguide connected to the SOA into a width W2 of the narrow waveguide region, and the passive waveguide makes an angle θwg with a normal to an end face of the substrate and is connected to the end face of the substrate.

[0011] As described above, according to one embodiment of the present invention, a light emitter having high manufacturing stability and high reflection suppression effect can be provided. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 This is a cross-sectional view schematically showing the structure of a general EADFB laser.

[0013] Figure 2 It is a cross-sectional view showing a schematic structure of AXEL.

[0014] Figure 3 This is a top view showing the schematic structure of the emission end face of AXEL.

[0015] Figure 4 It is a top view showing a plurality of semiconductor chips arranged on a wafer.

[0016] Figure 5 (a) and (b) are diagrams for explaining the chip yield during wafer manufacturing.

[0017] Figure 6 It is a top view showing a schematic structure near the emission portion of the AXEL having a window region.

[0018] Figure 7 The chart shows the results of evaluating the optical waveform quality of multiple modules equipped with AXEL.

[0019] Figure 8 is a diagram schematically showing the shape of a beam near the end face of a semiconductor chip. Figure 8 (a) is the top view, Figure 8 (b) is a cross-sectional view.

[0020] Fig. 9(a), (b) and (c) are cross-sectional views showing a part of the production process of the window region.

[0021] Fig.10 is a schematic diagram showing a schematic structure of an AXEL having a window region formed by embedding regrowth, Fig.10 (a) is the top view, Fig.10 (b), (c) and (d) are cross-sectional views.

[0022] Fig.11 The figure shows the results of evaluating the light waveform quality in the window region for a plurality of modules equipped with the AXEL having the window region.

[0023] Fig.12 This is a diagram showing the vicinity of the emission end of the rear stage of the SOA of AXEL in one embodiment of the present disclosure.

[0024] Fig.13 The figure shows a plurality of adjacent chips arranged on a wafer before cleavage.

[0025] Fig.14 is a graph showing the dependence of the mask margin on the width of the narrow waveguide region

[0026] Fig.15 This is a diagram showing the results of evaluating optical loss of a plurality of modules equipped with an AXEL having a narrow waveguide region according to an embodiment of the present disclosure.

[0027] Fig.16 The diagram is for explaining the miniaturization of the chip size and the improvement of the chip yield achieved by the AXEL having a narrow waveguide region in one embodiment of the present disclosure.

[0028] Fig.17 FIG. 1 is a top view showing a schematic structure of an AXEL according to an embodiment.

[0029] Fig.18 is a diagram schematically showing a configuration example of an AXEL chip used in one embodiment.

[0030] Fig.19 is a diagram showing the state of the end of a mesa-shaped waveguide during the AXEL manufacturing process. Fig.19 (a) is a cross-sectional view after dry etching. Fig.19 (b) and (c) are cross-sectional views after wet etching treatment.

[0031] Fig. 20 This is a diagram showing a schematic structure of an AXEL chip according to one embodiment.

[0032] Fig.21It is a diagram showing the relationship between the width W3 of the waveguide end and the side etching amount.

[0033] Fig. 22 This is a top view showing a schematic structure near the end of a waveguide in an AXEL chip of one embodiment. DETAILED DESCRIPTION

[0034] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. The same or similar reference numerals represent the same or similar elements, and repeated descriptions are sometimes omitted. The numerical values ​​or materials in the following description are illustrative, and other numerical values ​​or materials can be used to implement the present disclosure as long as they do not deviate from the main purpose.

[0035]

Formula 1

[0036] In addition, the substrate orientation

[011] in the following description may be the substrate orientation In the following description, only the substrate orientation

[011] is described, and Records of.

[0037] Before describing the embodiments of the present disclosure, AXEL, which is an EADFB laser integrated with an SOA, will be described.

[0038] Figure 1 This is a cross-sectional view schematically showing the structure of a general EADFB laser. Figure 1 The general EADFB laser 10 shown has a structure in which a DFB laser 11 and an EA modulator 12 are integrated into the same chip. The DFB laser 11 has an active layer 13 formed of a multi-quantum well (MQW), and oscillates at a single wavelength through a diffraction grating 14 formed in a resonator. In addition, the EA modulator 12 has a light absorption layer formed of a multi-quantum well (MQW) having a composition different from that of the DFB laser 11, and changes the amount of light absorption by voltage control. The EA modulator 12 is driven under the condition of transmitting / absorbing the output light from the DFB laser 11 to flicker the light, thereby converting an electrical signal into an optical signal.

[0039] The EADFB laser 10 has a problem that it is difficult to increase the output because of large optical loss in the EA modulator. The above-mentioned AXEL has been proposed as a solution to this problem.

[0040] Figure 2 It is a cross-sectional view showing a schematic structure of AXEL. Figure 2The AXEL20 shown modulates the light from the DFB laser 11 into signal light through the EA modulator 12, and amplifies the signal light through the SOA16, so that the output can be improved. AXEL can obtain about twice the high output characteristics compared with the general EADFB laser. In addition, AXEL can achieve high-efficiency operation through the SOA integration effect, so when driven under the working conditions that can obtain the same light output as the general EADFB laser, it can reduce the power consumption by about 40%. Moreover, in AXEL, the active layer 13 of the SOA16 and the DFB laser 11 is formed using the same MQW structure. Therefore, AXEL20 can be manufactured using the same manufacturing process as the previous EADFB laser 10 without adding the integrated regrowth process for SOA16.

[0041] AXEL has the following problem: the degradation of the operating characteristics caused by the reflected return light caused by the high light output characteristics associated with the integration of SOA16 becomes significant. In general optical transmitters such as semiconductor lasers, it is known that the reflected return light reflected from the end face of the semiconductor chip and returned to the inside of the chip has an adverse effect on the operating characteristics of the device. Therefore, in semiconductor optical transmitters, an anti-reflection film (Anti-Reflection (AR-) Coating) is implemented on the end face of the chip, and the reflected return light from the end face of the chip to the inside is generally suppressed to less than 0.1%. However, in the case of the EADFB laser (AXEL) integrated with SOA, due to its high output characteristics, even a tiny reflected return light will have a great impact on the operating characteristics. Assuming that the optical amplification effect achieved by the SOA in AXEL is +3dB compared to the previous EADFB laser, while the average optical output is increased by +3dB, the reflected return light intensity is also increased by 3dB. Furthermore, the reflected return light from the end face of AXEL is amplified again in the SOA, so the intensity of the reflected return light from the DFB laser reaching AXEL is increased by +6dB compared to the EADFB laser. Therefore, as a countermeasure for reflected return light in AXEL, in addition to the AR coating, a structure combining a window structure and a curved waveguide is adopted.

[0042] Figure 3 This is a top view showing the schematic structure of the emission end face of AXEL. Figure 3The figure shows a portion of a waveguide 32 connected to the SOA of the chip-based AXEL 20. The signal light from the SOA of the AXEL 20 in the semiconductor chip (also simply referred to as a chip in this specification) 31 is guided in the waveguide 32 and output from the output end face 33. Generally, the output end face of the semiconductor chip is a crystal plane formed by cleavage, and the waveguide in the semiconductor chip is formed at an angle perpendicular to the output end face. Therefore, the light transmitted in the waveguide is perpendicularly incident on the output face and is emitted from the semiconductor chip. In contrast, in Figure 3 In the AXEL 20 shown, a bend is provided in the waveguide 32 in the semiconductor chip 31. The signal light guided in the waveguide 32 is incident at an incident angle of θwg relative to the output end face 33. As a result, the light reflected at the output end face 33 is not easily coupled to the waveguide 32 again. In other words, the reflected return light from the output end face 33 can be suppressed. Generally, the incident angle θwg of the signal light from the waveguide 32 relative to the output end face 33 for suppressing reflection is 4 to 8 degrees.

[0043] Moreover, in Figure 3 In the structure shown, the waveguide 32 terminates inside the semiconductor chip 31. The signal light emitted from the waveguide 32 is transmitted in the bulk semiconductor called the window area 35, reaches the output end face 33, and is emitted from the semiconductor chip 31 to the outside. At this time, due to the diffraction effect in the window area 35, the signal light is transmitted while expanding the beam diameter. As a result, the ratio of the light field reflected by the output end face 33 and overlapping with the waveguide 32 inside the semiconductor chip 31 becomes smaller, so the reflected return light coupled with the waveguide 32 can be further reduced. Generally speaking, the length of the light transmission direction (X direction) of the window area 35 is made to be about 10μm. In AXEL20, if these reflected return light countermeasures are not fully implemented, it is impossible to obtain high-quality transmission characteristics while obtaining high output characteristics.

[0044] As described above, in AXEL, in order to minimize the reflected return light at the end face of the chip, a curved waveguide and a window structure are used at the output end face. Here, the problem of this structure is explained.

[0045] First, refer to Figure 4 , the problems for the case of using curved waveguides are described. Figure 4 3 is a top view showing a plurality of semiconductor chips 31 arranged in the X direction and the Y direction in the wafer. It should be noted that in this specification, the semiconductor substrate that forms the AXEL and is chipped is also referred to as the AXEL chip. Figure 4 In the embodiment, two semiconductor chips 31 each having an AXEL having a waveguide 32 are opposed to each other in the light transmission direction (X direction). The waveguide 32 includes a straight waveguide portion 32a and a curved waveguide portion 312b. Figure 4An AXEL chip without the window region 35 is shown.

[0046] Generally, for optical semiconductor devices such as semiconductor lasers, multiple chips having the same output end face are cleaved in strips to form output end faces. In addition, adjacent bars (BARs) in the light transmission direction are arranged in a manner that the output end faces face each other, and the opposing chips are arranged in a manner that they share the output end face formed by cleavage.

[0047] like Figure 4 As shown, a plurality of chips 31 form a curved waveguide portion 32b in a manner that the optical axis is θwg relative to the emission end face, and are configured in a state where the curved waveguides 32b in two chips 31 adjacent and opposite to each other in the X direction are connected to each other. That is, the two adjacent and opposite chips are configured in a manner that the optical axes are the same. In other words, the two chips are continuously formed on the same waveguide. The two adjacent bars 1 (Bar1) and 2 (Bar2) are cut apart by cleavage to form an emission end face 33 where the waveguide 32 terminates. In the actual cleavage process of the semiconductor chip, the accuracy of the cleavage position is limited, and a positional offset of about ±10μm usually occurs. As described above, by aligning the optical axes of the opposing chips and connecting the waveguides, even if an error occurs in the cleavage position, a structure in which the waveguide terminates at the emission end face can be formed in the two chips, thereby suppressing the reduction in yield in the cleavage process. However, this configuration will result in the disadvantage of large chip size.

[0048] The DFB laser 11, EA modulator 12, and SOA 16 in the AXEL are formed on a waveguide having an optical axis in a direction (X direction) perpendicular to the emission end face 33. Here, in the two opposing chips, the optical axis of the DFB laser 11 is not aligned, but is arranged in parallel at a distance ΔY apart. This ΔY is called a waveguide offset. Figure 4 It can be seen that the dimension of the semiconductor chip 31 in the Y direction needs to be larger than the dimension of the waveguide offset. In addition, ideally, the light emission position of the optical element part such as the DFB laser 11 and the cleavage end face (the termination position of the waveguide 32) is located in the center of the chip. Therefore, the dimension of the chip in the Y direction is designed to have a sufficient margin with respect to the waveguide offset. Here, the curved waveguide portion 32b of a conventional AXEL requires a minimum length of about 400μm. In addition, assuming that the bending angle θwg of the waveguide is 5°, the waveguide offset is about ΔY=70μm. In this case, while ensuring a margin of 100μm on both sides of the straight waveguide portion 32a of each chip 31 in bar 1 and bar 2 ( Figure 4In order to ensure the margin 1 and margin 2 in the chip, a minimum chip width (Y-direction length) of 270 μm is required. Thus, in the previous AXEL, it was difficult to miniaturize the width of the AXEL chip (dimension in the Y direction) to less than 250 μm in order to ensure sufficient margin (distance) relative to the waveguide position in the chip from the perspective of manufacturing margin and long-term reliability. In addition, in the previous AXEL configuration, the chip yield during wafer manufacturing also became a problem.

[0049] Reference Figure 5 , explaining the chip yield during wafer manufacturing. Figure 5 (a) schematically shows the configuration of four chips (chips 1 to 4) continuously formed on the same optical waveguide 32 on a semiconductor wafer during the manufacturing process of the AXEL chip. As described above, after the manufacturing process of each element in the chip, each chip is separated by a cleavage process. Chip 1 and chip 2 are arranged / connected opposite to each other in such a manner that SOA 16 for obtaining light output is close, and chip 2 and chip 3 are connected in such a manner that the DFB laser is adjacent. As described above, the output end face of the chip and the end of the waveguide are determined by cleavage. Therefore, the position offset error of the cleavage is taken into account in advance, and the chips adjacent in the x direction are configured in a state of connecting the waveguide 32. This applies not only to the SOA side (the boundary between chip 1 and chip 2) which is the output end side, but also to the DFB laser side (the boundary between chip 2 and chip 3).

[0050] Here, due to the aforementioned waveguide offset, a positional offset in the Y direction occurs between chip 2 and chip 3. Figure 5 (a) of FIG. 3 shows only four chips arranged continuously in the X direction, but on an actual wafer, more than ten chips are formed on the same continuous waveguide 32. Therefore, the more chips there are in a row, the more the positional offset in the Y direction caused by the waveguide offset will accumulate. When ten chips are uniformly arranged on a wafer, if the waveguide offset ΔY is set to 40 μm, the Y-direction offset between chip 1 and chip 10 reaches 400 μm. Such positional offset on the wafer will have an adverse effect on the chip configuration.

[0051] Figure 5 (b) is a schematic diagram of manufacturing when 10×10 AXEL chips are arranged on a wafer. Generally, in the manufacturing process of optical semiconductors, a basic unit called a shot area (shot) is divided and arranged on a wafer. Figure 5 As shown in (b), when a waveguide offset occurs, a dead space (blank) is created where chips cannot be placed in the irradiation area, which inevitably leads to a reduction in the number of chips manufactured on the same wafer and a reduction in yield. Figure 5In (b), if the length of the AXEL chip (length in the X direction) is set to 1000 μm, the chip width (length in the Y direction) is set to 300 μm, and the waveguide offset is set to 70 μm, the dead space in the irradiation area is 700,000 μm. 2 The size of the AXEL chip is 300,000 μm. 2 , which will result in a decrease in overall yield of about 2 to 3%.

[0052] Next, refer to Figure 6 A problem in the case of using a window area is described. Figure 6 It is a top view showing a schematic structure near the emission portion of the AXEL having a window region. Figure 6 In order to make the configuration within the wafer clearly visible, adjacent chips within the same stripe (strip 1) and adjacent chips within the opposite stripe (strip 2) are shown. Figure 4 As stated in Figure 6 In the AXEL having the window region, the optical axis of the waveguide 32 at the output end is also configured to coincide with the optical axis of the waveguide 32 of the opposing chip.

[0053] In addition, the two chips facing each other each have a window region 35, and the end of the waveguide 32 is located at a position separated from the predetermined cleavage position 44 by the length of the window region 35 (length in the X direction). The length of the window region is generally designed to be about 10 μm. Here, as described above, a positional deviation error of about ±10 μm occurs during the cleavage process. Therefore, when the error of the cleavage position is greater than 10 μm, Figure 6 The window area 35 in any of the two stripes shown will disappear, and a sufficient reflection suppression effect cannot be obtained.

[0054] Figure 7 The results of evaluating the optical waveform quality of multiple modules equipped with AXEL manufactured in the same process are shown. Here, the AXEL chips are all chips with the same structure with an oscillation wavelength in the 1.3μm band, and the only difference between the chips is the difference in the length of the window area caused by manufacturing errors. After measuring the length of the window area of ​​each chip in advance, each chip was installed in a general butterfly-shaped semiconductor module composed of a dual-lens system. In addition, a high-frequency connector is also installed in the butterfly-shaped package, and each module is manufactured with 25Gbit / s NRZ signal modulation, and the optical waveform (EYE waveform) is evaluated. Figure 7 The vertical axis is an index called mask margin that represents the quality of the light waveform. The larger the margin, the clearer the eye opening. Figure 7The horizontal axis represents the deviation between the length of the window area of ​​each manufactured AXEL chip and the design value when the design value of the window area length is set to 10μm. As a trend, it can be confirmed that the smaller the length of the window area, the worse the mask margin. It is believed that this is because as the length of the window area decreases, the intensity of the return light reflected from the chip end face to the inside of the chip increases, causing the operation of the AXEL to be unstable. Figure 7 It can be seen that the minimum length of the window region needs to be 5 μm or more. In addition, if the window region 35 is too long, it will also have an adverse effect on the characteristics of AXEL.

[0055] Figure 8 1 is a diagram schematically showing a beam shape near an end face when the length L of the window region 35 becomes longer than designed due to a manufacturing error of the cleavage position 44 . Figure 8 (a) is the top view, Figure 8 (b) is a cross-sectional view. In the case of the communication bands of 1.55μm band and 1.3μm band, the thickness of the core layer is about 200nm to 300nm. Generally, in InP-based optical semiconductor devices, light confinement in the vertical direction (Z direction) of the waveguide is stronger than that in the horizontal direction (XY direction). Therefore, the light beam divergence angle in the vertical direction in the window area 35 is larger than the light beam divergence angle in the horizontal direction. Therefore, when the window area is long, the upper end of the light beam reaches the interface between the cladding and the outside of the semiconductor (air, electrode) in the window area, which will cause defects (blocking) in the beam shape and light loss ( Figure 8 (b)). In general InP-based optical semiconductor devices, the cladding layer on the waveguide is formed by regrowth, and the thickness of the upper cladding layer is about 2μm. When a window structure is set, the cladding layer thickness on the core layer needs to be designed in consideration of the beam divergence due to the diffraction effect of the beam.

[0056] The InP thickness of the window region 35 is likely to be thinner than the thickness of the upper cladding layer of the waveguide part, so it is particularly likely to affect the optical field. Fig. 9 The reason is explained. As described above, in the usual window region 35, the waveguide ends and becomes an InP region. The InP of the window region 35 is formed simultaneously during the upper cladding growth process or the embedded growth process. Here, the case where the window region is formed during the embedded growth is explained. Fig. 9 FIG. 1 is a cross-sectional view showing a part of the manufacturing process of an AXEL having an embedded heterostructure (BH structure). The DFB laser 11, the EA modulator 12, and the SOA 16 have a common embedded structure waveguide. Fig. 9As shown in (a), first, an insulating film is formed on the surface of the lower cladding layer, the core layer 83, and the upper cladding layer 84 stacked on the InP substrate 81, and then an insulating layer mask 85 in a waveguide shape is formed by patterning. Then, as shown in FIG. Fig. 9 As shown in (b), a mesa forming step is performed by a dry etching process or a wet etching process to form a waveguide including a waveguide core 83a. Fig. 9 As shown in (c), the waveguide is embedded by regrowing the InP layer 86 as a current blocking layer, and then the insulating layer mask 85 is removed to complete the BH structure. The aforementioned window region 35 is also formed at the same time when the BH structure is manufactured.

[0057] Fig.10 Schematic diagram showing a schematic structure of an AXEL having a window region formed by embedded regrowth. Fig.10 (a) is a schematic diagram of the AXEL window region as viewed from the upper surface of the substrate.

[0058] Fig.10 (b) is a cross-sectional view of the AXEL waveguide. Fig.10 (c) is a cross-sectional view of the optical waveguide. Fig.10 (d) is a cross-sectional view of the window region. Here, a straight waveguide without a curved waveguide is illustrated for simplicity of description. Fig.10 As shown in (b), InP86 as a current blocking layer is formed by embedding and regrowing on both sides of the mesa-shaped waveguide, and the waveguide core 83a becomes a structure in which InP86 is embedded. Fig.10 As shown in (c) and (d), the portion that becomes the window region when the mesa shape of the waveguide is produced is also etched at the same time, and an InP layer identical to the current blocking layer InP is embedded in the embedding and re-growth process.

[0059] In conventional embedding regrowth, the thickness of the InP layer to be regrown is set according to the height of the mesa-shaped waveguide so that the waveguide is completely embedded. This is because if the InP layer to be embedded regrowth is too thick compared to the waveguide, the InP layer will climb up and accumulate on the Fig. 9 The upper part of the insulating layer mask 85 shown in (c) is shown. Fig.10 As shown in the top view of (a), the window region is the part where the mesa shape ends, so the opening area embedded during regrowth is relatively wide compared to other parts. Therefore, when the mesa-shaped waveguide and the window region are embedded with InP at the same time by regrowth, as shown in FIG. Fig.10As shown in (c), the thickness of InP in the window region is thinner than that of the mesa-shaped waveguide. As described above, in the window region, due to the diffraction phenomenon, the light field is transmitted while being divergent. Therefore, when the thickness of the InP film is thin, the light field contacts the upper air interface, which easily causes light loss.

[0060] Fig.11 This is a graph showing the results of manufacturing a plurality of modules equipped with AXELs having window regions manufactured using a conventional embedding regrowth process and evaluating the relationship between the optical loss in the module and the fluctuation in the length of the window region. Figure 7 The data shown are the results of evaluating the same modules. As mentioned above, AXEL chips with the same structure are used in each module, and the design value of the length of the window area is 10μm. Before installing each AXEL chip in the module, the light output characteristics are evaluated using a large-aperture photodetector. Then, after the AXEL chip is installed in the module, it is coupled to the optical fiber and the light intensity is evaluated to estimate the light loss caused by installation in the module. Here, the module used has a dual-lens system, and installation in the module is performed through an active alignment process. From Fig.11 It can be seen that if the length of the window area becomes longer, the loss inside the module tends to increase. This is because, that is, as the length of the window area becomes longer, the upper end of the outgoing light beam reaches the boundary between the cladding and the outside of the chip, and the shape of the light beam is imperfect, thereby reducing the coupling efficiency to the optical fiber. Here, in order to suppress the optical loss in the module to less than 3.0 dB and produce a stable optical module, the window area length of the AXEL chip needs to be less than 15 μm, that is, the offset of the cleavage position is less than 5 μm. However, in the actual cleavage process of the semiconductor chip, a cleavage position error of about ±10 μm is generally generated. In the AXEL with a window structure, in order to manufacture an optical transmitter with sufficient characteristics, it is impossible to set a sufficient margin when forming the output end face by cleavage. Therefore, due to the manufacturing error of the cleavage position, a certain number of defective products (chips whose window area length deviates from the allowable value) will inevitably be produced, which becomes the reason for reducing the manufacturing yield.

[0061] According to one embodiment of the present disclosure, an AXEL, that is, a light emitter, having high manufacturing stability and high reflection suppression effect can be provided. Hereinafter, an AXEL according to one embodiment of the present disclosure will be described.

[0062] (Implementation Method)

[0063] As described above, in the window area, the light beam is transmitted while being divergent due to diffraction. Therefore, even in the case of being reflected by the end face, the ratio of the light field distribution that is re-coupled with the waveguide can be reduced, thereby reducing the amount of returned light. Therefore, in the present disclosure, the shape of the light beam is enlarged by narrowing the waveguide width using a taper. In principle, when the width of the waveguide is sufficiently small, it will be cut off, and the light will not be trapped in the waveguide and radiate. In this state, the light beam is transmitted while being divergent, so the same reflection suppression effect as that of the part in the window area can be obtained. However, it is difficult to stably manufacture a narrow waveguide in which the transmitted light is cut off in the communication band (1.3μm or 1.55μm). In the present disclosure, it is not necessary to use a narrow waveguide that meets the cutoff condition, and the full effect can be exerted by making the waveguide width so that the beam diameter of the waveguide light is sufficiently large within the range of the manufacturing margin.

[0064] Fig.12 This situation is shown. Fig.12 This is a diagram showing the vicinity of the output end of the rear section of the SOA of AXEL according to an embodiment of the present disclosure. In the present disclosure, a tapered region 32c and a narrow waveguide region 34d for narrowing the waveguide width are provided near the output end of the curved waveguide portion 32b of the waveguide 32 that guides the light from the SOA. The narrow waveguide region 32d has a fixed length and is configured to exceed the designed predetermined cleavage position 44. Here, the narrow waveguide region 32d has a sufficient length for the positional offset error of the cleavage process. Therefore, even in the case of a cleavage position offset, the end face will definitely be formed in the narrow waveguide region 32d, so that the quality of the output light beam will not be affected. In addition, the waveguide end 32e is configured on the outside of the predetermined cleavage position, that is, outside the chip. With such a configuration, referring to Fig.10 , the transmission light will not be affected by the thinning of the InP86 film thickness and the deterioration of the surface flatness in the window region at the waveguide end 32e as described above.

[0065] Here, use Fig.13 The length required for the narrow waveguide region is explained. In the cleavage process described above, a positional deviation error of about ±10 μm will occur. Therefore, the narrow waveguide region 32d needs to have a sufficient length of more than 20 μm to compensate for this error ( Fig.13 In addition, in order to fully obtain the reflection suppression effect, the narrow waveguide region 32d after the tapered region 32c needs to have a minimum of 5 μm or more of residual ( Fig.13 Furthermore, it is necessary to ensure that the residual waveguide 32f remaining in the adjacent chip when the cleavage position shift occurs is about 5μm ( Fig.13M3). Considering the above aspects, the minimum length of the narrow waveguide region 32d needs to be about 30 μm. In this case, if the above-mentioned error in the cleavage position is taken into account, the length of the narrow waveguide region 32d in the cleaved AXEL chip is 5 μm to 25 μm, and the length of the residual waveguide 32f located in the adjacent chip and including the waveguide end 32e is 5 μm to 15 μm.

[0066] Furthermore, by adjusting the arrangement of the waveguide 32 between adjacent chips, an arrangement that eliminates the waveguide deviation can be achieved. Fig.13 It represents a plurality of adjacent chips arranged on a wafer before cleavage. In the present disclosure, the waveguide end 32e is arranged outside the chip. That is, the waveguide end 32e is arranged inside an adjacent chip facing each other. Therefore, the structure is as follows: by utilizing the waveguide offset, the position of the waveguide end 32e is arranged at a position sufficiently away from the optical axis transmitted in each other's chips, so that the influence of the waveguide end 32e does not affect the guided light. Therefore, the waveguide end 32e of the residual waveguide 32f of the adjacent chip on the wafer remains inside the chip after cleavage. The waveguide end 32e of the residual waveguide 32f also does not affect the chip operation at all.

[0067] In addition, by adopting Fig.13 The arrangement shown alleviates the limitation of chip size caused by waveguide offset. That is, the waveguides are arranged so that the optical axes of the chips facing each other do not coincide with each other, so the chip size can be shortened.

[0068] By using the aforementioned narrow waveguide region 32d, a sufficient reflection suppression effect can be obtained. Here, firstly, AXEL chips with different widths W perpendicular to the light propagation direction in the narrow waveguide region 32d are manufactured to evaluate the influence of reflected return light. Figure 7 As discussed in the previous section, a butterfly module with a built-in 1.3μm-band AXEL was fabricated, and the impact of the return light quantity was evaluated based on the waveform quality of the 25Gbit / s optical waveform. The evaluation used AXEL chips with narrow waveguide regions of different widths W. AXEL chips with the same structure except for the width W of the narrow waveguide region were used.

[0069] Fig.14 It is a graph showing the dependence of the width of the narrow waveguide area of ​​the evaluated mask margin. The mask margin is an indicator showing the waveform quality of the transmitted light. The larger the value of the mask margin, the higher the waveform quality of the transmitted light. The width W of the narrow waveguide area of ​​the mesa shape was varied from 0.25μm to 1.5μm for evaluation. Here, W=1.5μm refers to a chip in which the width of the waveguide of a conventional AXEL is fixed, that is, a chip in which a waveguide of the same width is not used near the emission end of the waveguide, but has a waveguide of the same width up to the end face. By Fig.14It can be seen that the narrower the width W of the narrow waveguide region at the output end, the more the end face reflection is suppressed, and the more the waveform quality can be improved. If the width W of the narrow waveguide region is less than 0.7μm, the mask margin can be guaranteed to be more than 30%, which can be said to have a sufficient reflection suppression effect in practical use. However, when the width W of the narrow waveguide region is 0.25μm, it is difficult to stably manufacture in the photolithography process, which is a general manufacturing process for optical semiconductor devices. As a structure that takes into account both manufacturing stability and sufficient reflection suppression effect, the width W of the narrow waveguide region is about 0.4μm to 0.8μm, which can best exert the effect.

[0070] Next, the influence of the shift of the cleavage position in the AXEL having a narrow waveguide region of the present disclosure on the chip characteristics was evaluated. Fig.15 The results are obtained by evaluating the optical loss of multiple modules equipped with AXEL chips produced in a conventional cleavage process. In the AXEL chip here, the width of the narrow waveguide region is set to W = 0.5 μm. Fig.15 The data shown in the previous AXEL is Fig.11 In addition, the evaluation is the same as Fig.11 The content described in is the same as that of the calculation of the optical loss in the module based on the change in the optical output before and after the module is modularized. If the beam is blocked at the top of the cladding, the loss may increase. Fig.15 In the conventional AXEL chip, as the cleavage position shifts, the optical loss increases. In contrast, in the AXEL with a narrow waveguide region disclosed in the present invention, the optical loss does not depend on the cleavage position shift and becomes a roughly constant value, which is roughly 2.5 dB lower than the loss. From this result, it can be confirmed that by introducing the narrow waveguide region disclosed in the present invention, the fluctuation of chip performance caused by the shift of the cleavage position, which has been a problem in the past, is greatly suppressed.

[0071] In addition, as an effect of the AXEL having a narrow waveguide region of the present disclosure, it is possible to list the miniaturization of the chip size and the improvement of the chip yield during wafer manufacturing. Fig.16 This figure illustrates the miniaturization of chip size and the improvement of chip yield achieved by AXEL with a narrow waveguide region. Fig.16As shown, the straight waveguide portion 32a of the waveguide 32 of AXEL can be arranged in the center of the chip in the Y direction. Therefore, even if the length of the chip in the Y direction is reduced, sufficient margin can be ensured on both sides of the waveguide. Here, by setting the size of the chip in the Y direction to 200μm and arranging the straight waveguide portion 32a in the center of the chip, a sufficient margin of 100μm is ensured on both sides of the upper and lower sides of the waveguide. When the angle θwg formed by the optical axis of the curved waveguide portion 32b and the normal (X direction) of the cleavage surface (end face) is set to 5° and the length of the curved waveguide portion 32b is set to 400μm, the positional offset of the chip output end is about 35μm, which is a configuration that is fully allowed in manufacturing.

[0072] In addition, Fig.16 In the embodiment, the optical axes of the linear waveguide portions 32a of the adjacent chips are aligned. Thus, the waveguide offset ΔY in the Y direction can be set to 0. Figure 5 The problem of chip yield caused by waveguide offset as described above. Here, in a general AXEL chip, Fig.16 The length of the curved waveguide portion 32b is about 150 μm to 500 μm. In addition, the angle θwg formed by the optical axis of the curved waveguide portion 32b and the normal to the cleavage surface (end face) is generally 4° to 8° as described above. Therefore, when the angle θwg is assumed to be 4°, Fig.16 The offset D / 2 of the emission position in is 10μm<D / 2<35μm. In addition, when the angle θwg is assumed to be 8°, it is 21μm<D / 2<70μm. Therefore, the narrow waveguide region 32d of the adjacent AXEL chip is separated by a distance D on the cleavage plane, and D is 10μm<D<140μm. However, in actual use, the surface flatness of the waveguide end 32e decreases during regrowth as described above, so it is necessary to maintain a minimum distance of 30μm from the curved waveguide portion 32b that guides light, so the ideal is 30μm<D<140μm.

[0073] (Example 1)

[0074] Reference Fig.17 The AXEL of Example 1 is described. This example is an optical transmitter capable of generating a 25 Gbit / s modulated signal, which realizes an optical transmitter that increases the optical output during modulation to 9 dBm or more in order to cope with a system with a high loss budget.

[0075] Fig.17 It is a top view showing the schematic structure of AXEL according to this embodiment. Fig.17The semiconductor chip 31 shown has a waveguide 32, a DFB laser 11, an EA modulator 12, and an SOA 16 connected by the waveguide 32. As described above, the waveguide 32 has a curved waveguide portion. The curved waveguide portion has a tapered region and a narrow waveguide region. The semiconductor chip 31 is an AXEL chip that monolithically integrates the DFB laser 11, the EA modulator 12, and the SOA 16, and constitutes an optical transmitter. Fig.17 Another AXEL chip (semiconductor chip 31) adjacent to the wafer is also shown in the figure to facilitate understanding of the process of forming the emission end face by cleavage.

[0076] In the optical transmitter (semiconductor chip 31) of the present embodiment, the length of the DFB laser 11 (the length along the waveguide 32, the same below) is 300 μm, the length of the EA modulator 12 is 150 μm, and the length of the SOA 16 is 200 μm. In addition, the waveguide 32 adopts an embedded heterostructure using semi-insulating InP that can obtain high heat dissipation effect and current narrowing effect. As described above, in order to obtain a sufficient reflection suppression effect, the AXEL chip of the present embodiment is provided with a tapered region 32c and a narrow waveguide region 32d in the curved waveguide portion 32b. The narrow waveguide region 32d is located near the emission end face. The AXEL chip is formed on the InP substrate (100) surface, and the DFB laser 11 is configured to output light in the direction of the substrate orientation

[011] . After passing through the EA modulator 12 on the same optical axis as the DFB laser 11, the light from the DFB laser 11 changes its transmission direction in the curved waveguide portion so that it has an angle θwg relative to the crystal orientation

[011] and is incident on the SOA 16. The light emitted from the SOA 16 maintains the angle θwg and is converted to the width W of the narrow waveguide region in the tapered region. The light transmitted to the narrow waveguide region ends inside the adjacent chip 31 after passing through the predetermined cleavage position 44. Here, as a bending angle that can obtain a sufficient reflection suppression effect, the angle θwg is set to 5°. In addition, in order to obtain a sufficient reflection suppression effect, the width W of the narrow waveguide region, which is a range that can be stably manufactured in a semiconductor processing process, is 0.5μm.

[0077] Under this condition, the narrow waveguide region does not reach the cutoff condition. Therefore, light is transmitted in the narrow waveguide region without being radiated. Moreover, the two AXEL chips adjacent to each other in the X direction are opposite to each other. In addition, in the two AXEL chips, the straight waveguide parts in which the DFB laser 11 and the EA modulator 12 are formed are arranged in parallel and with the optical axes aligned. The curved waveguide parts of the two AXEL chips are arranged in parallel and are separated by a distance D at the predetermined cleavage position. Here, the length L_bend of the curved waveguide part of the AXEL chip is 400μm. Therefore, the waveguide distance D is 70μm. The two AXEL chips are separated by the cleavage process. As mentioned above, the cleavage process usually produces a positional offset error of about ±10μm. Therefore, the light emission position of the final chip is determined at the moment when the cleavage process is completed.

[0078] In order to obtain sufficient reflection suppression effect, it is necessary to ensure that the length of the narrow waveguide area after the tapered area in the AXEL chip is more than 5 μm ( Fig.13 In addition, considering the positional deviation error in the cleavage process, it is necessary to ensure that the cleavage error compensation area L_scr is greater than 20μm ( Fig.13 M1), the length of the residual waveguide in the adjacent AXEL chip is guaranteed to be more than 5μm ( Fig.13 M3). Therefore, in this embodiment, the length L_nar of the narrow waveguide is designed to be 30μm. When the position offset of the cleavage process is set to 0, a 15μm narrow waveguide portion remains in the AXEL chip. In addition, even if the cleavage position offset occurs and the narrow waveguide area becomes shorter, as long as the cleavage error is within the range of ±10μm, a narrow waveguide area of ​​more than 5μm will remain in the chip. On the contrary, even if the narrow waveguide area becomes longer due to the cleavage position offset, the length of the narrow waveguide area in the chip is less than 25μm, and the length of the residual waveguide remaining in the oppositely configured AXEL chip must be more than 5μm.

[0079] Next, the manufacturing process of the light emitter (semiconductor chip 31) of this embodiment is described. In the manufacture of the semiconductor chip 31, an initial substrate is used in which a lower SCH (Separated Confinement Heterostructure) layer, an active layer (MQW1) of a multi-quantum well layer, and an upper SCH layer are sequentially grown on the surface of an n-InP substrate (100). The multi-quantum well layer is configured to have optical gain in the oscillation wavelength 1.3μm band. Here, the multi-quantum well includes six quantum well layers. The initial substrate is a structure optimized for high-efficiency operation of the DFB laser 11.

[0080] First, the portion that becomes the DFB laser 11 and SOA16 is left, and the other active layers are selectively etched, and then a multi-quantum well layer (MQW2) is grown on the portion that becomes the EA modulator 12 by butt-jointing regrowth. Here, the configuration is to transmit light in the order of the DFB laser 11, the EA modulator 12, and the SOA16. In addition, the SOA16 directly utilizes the layer structure of the core layer formed on the initial substrate. Therefore, the layer structure of the SOA16 is the same as that of the DFB laser 11. Ultimately, the difference in the layer structure between the DFB laser 11 and the SOA16 is only the presence or absence of a diffraction grating. Thus, it is a structure that integrates multiple regions, and the number of regrowths can be suppressed, and it can be manufactured at a low cost.

[0081] Next, the region that becomes the boundary of the DFB laser 11 and the EA modulator 12, the region that becomes the boundary of the EA modulator 12 and the SOA 16, and the region from the end of the region that becomes the SOA 16 to the emission end face are selectively etched again, and a bulk semiconductor that becomes the core layer of the passive region is grown by butt-joint regrowth. The core layer of the passive region has a band gap wavelength shorter than the oscillation wavelength of the DFB laser. In this specification, the waveguide including the passive region connected to the end of the region that becomes the SOA 16 is also referred to as a passive waveguide.

[0082] Next, a diffraction grating 14 (not shown) is formed on the portion that becomes the DFB laser 11 so as to operate in the oscillation wavelength 1.3μm band. Here, the diffraction grating 14 is configured so that the resonator of the DFB laser 11 outputs light in the direction of the substrate orientation

[011] . Then, the p-InP cladding layer and the contact layer are grown again on the entire surface of the semiconductor chip 31 by regrowth. Next, a mesa structure is formed on the portion that becomes the waveguide 32. Here, the waveguide width of the DFB laser 11, the EA modulator 12, and the SOA16 is 1.5μm, the same as that of general semiconductor devices. In this process, the tapered area and narrow waveguide area of ​​the curved waveguide portion and the waveguide end are also made. An insulating layer mask 85 is formed on the portion that becomes the waveguide 32, and a mesa structure is formed by dry etching.

[0083] Next, the mesa structure is embedded by embedding the regrown InP86 while keeping the insulating layer mask remaining. The embedding regrowth adjusts the growth amount of InP so that the mesa structure waveguide 32 is completely embedded. In the case where the growth amount of the semi-insulating InP is too small, the current blocking effect cannot be fully exerted, the current cannot be efficiently injected into the DFB laser 11 and SOA16, and sufficient characteristics cannot be obtained. In addition, in the case where the growth amount of the semi-insulating InP is too much, abnormal growth of the regrown InP climbing onto the insulating film on the upper part of the waveguide 32 will occur, and the electrode formation of the DFB laser 11 and SOA16 becomes difficult. Therefore, in order to ensure a sufficient current constriction effect and flatness of the upper part of the waveguide, the growth amount of InP during the embedding growth is adjusted so that the mesa structure waveguide 32 is completely embedded and the semi-insulating InP and the mesa structure waveguide 32 have the same height in the waveguide cross section. In this process, in the normal window area, the InP film thickness is thinned at the waveguide end 32e, which affects the light field (refer to Fig.10 (c) and (d)). However, in the present disclosure, since the waveguide end 32e is arranged outside the chip (adjacent chip), the light transmitted in the chip is not affected by the waveguide end.

[0084] Next, after removing the insulating layer mask 85, the contact layer between the regions is removed by wet etching in order to electrically separate the respective regions of the DFB laser 11, the EA modulator 12, and the SOA 16. Next, an electrode on the P side is formed for injecting current through the contact layer on each region on the upper surface of the semiconductor substrate. Then, the InP substrate 81 is polished to about 150 μm, and an electrode is formed on the back of the substrate, and the process on the semiconductor wafer is completed.

[0085] Next, a (011) crystal plane is formed by cleavage, thereby manufacturing a semiconductor strip including a plurality of chips 31. Here, a general semiconductor chip cleavage process is used, and the cleavage position accuracy is ±10 μm or less as described above. Using a semiconductor strip in which a plurality of chips are connected in the Y direction by the cleavage process, an AR coating is applied to the emission end face, and a high reflection coating (HR) is applied to the end face on the DFB laser 11 side.

[0086] In order to confirm the effect of this embodiment, twenty AXEL chips manufactured by the same process were installed in the module and evaluated. First, the optical output based on the large-aperture PD was evaluated in the AXEL chip before it was installed in the module. The input currents to the DFB laser 11 and SOA16 were set to 80mA and 40mA respectively, and the applied voltage of 0.6V and the electrical signal of 25Gbit / s were input to the EA modulator 12 for evaluation. As a result, the net optical output of the manufactured AXEL chip averaged about 12dBm in the output Pavg during modulation. Then, the optical output after installation in the module was evaluated. The module manufacturing process was installed based on the aforementioned active alignment process. As a result, it was confirmed that all twenty modules exceeded 9dBm in the optical output Pavg during modulation during optical coupling. Fig.15 The same evaluation as shown above shows that the loss in the module obtained by comparing the optical output intensity before and after installation in the module is less than 2.5dB, which can greatly improve the module manufacturing yield compared to the previous AXEL chip without a window area. Assuming that the previous AXEL chip is installed in the module, it is estimated that there will be a loss of more than 3dB when installed in the module, so it is difficult to manufacture an optical transmitter that achieves the target +9dBm optical output. In the AXEL chip of the previous technology, the yield of the chip manufactured in the wafer surface during modularization is about 40%, while in the AXEL module of this embodiment, the module manufacturing yield is improved to 65%.

[0087] Next, the module equipped with the AXEL chip of this embodiment was used to evaluate the modulation characteristics of 25 Gbit / s and confirm the working quality. The modulation signal used NRZ, PRBS2 31 -1. The driving conditions of AXEL are the same as those mentioned above. Here, the quality of the EYE waveform at 25Gbit / s was evaluated, and the average dynamic extinction ratio was 9.1dB. In addition, the mask margin calculated based on the EYE waveform was more than 30% in all modules. In this embodiment, the width of the narrow waveguide area was set to 0.5μm, and the above results confirmed Fig.14 In addition, according to these results, in the optical transmitter formed by the module of this embodiment, a stable mask margin of >30% can be obtained regardless of the deviation of the cleavage position, which is the same level as the good quality module obtained in the conventional AXEL chip. In the module equipped with the AXEL chip of this embodiment, only the manufacturing yield is greatly increased without deterioration of the waveform quality of the optical signal.

[0088] Finally, the transmission characteristics of a 40 km single-mode fiber were evaluated using the above working conditions, and it was confirmed that the bit error rate was less than 10 －12The above results confirm that the AXEL, which introduces a dummy mesa structure in the window area, dramatically improves the module manufacturing yield without degrading the performance of conventional modules.

[0089] (Example 2)

[0090] Reference Fig.18 The AXEL of Example 2 is described. In this example, the AXEL chip of the 1.5 μm band adopts a chip design that can increase the production volume within the same wafer surface. Fig.18 FIG. 1 is a diagram schematically showing a configuration example of the AXEL chip used in this embodiment.

[0091] Reference Figure 4 When two AXWL chips (two semiconductor chips 31) facing each other on the wafer are arranged so that the optical axes near the emission part are consistent (waveguide is continuous) as in the previous AXEL described above, the Y-direction positions of the DFB laser 11 and the EA modulator 12 having the waveguide 32 in the direction of the substrate crystal orientation

[011] are inconsistent in the two AXELs, resulting in a large offset. From the perspective of ensuring the margin during cleavage of the chip sidewalls and sufficient area of ​​the electrode, it is ideal that the DFB laser 11 and the EA modulator 12 are arranged at the center of the chip away from the chip sidewalls. Therefore, in the case of such a chip configuration, the chip width has to be designed to be larger. In a general AXEL chip, when such a configuration is adopted, the chip width (length in the Y direction) needs to be at least about 300μm.

[0092] In contrast, in the AXEL of this embodiment, Fig.18 As shown, the waveguides 32 of the two opposing AXEL chips are arranged on the wafer in such a manner that the optical axes of the curved waveguide portions 32b are not aligned, and the optical axes of the straight waveguide portions 32a of the waveguide 32 are aligned. Thus, the waveguides of the DFB laser 11 and the EA modulator 12 in the direction of the substrate crystal orientation

[011] can be arranged near the center of the AXEL chip, thereby reducing the chip width.

[0093] The manufacturing process of the AXEL chip of this embodiment is roughly the same as the process of the AXEL chip of the aforementioned embodiment 1. The lengths of the DFB laser 11, EA modulator 12 and SOA16 in the AXEL chip are 350μm, 200μm and 300μm, respectively. The diffraction grating 14 (not shown) formed in the DFB laser 11 has a period designed to make the oscillation wavelength 1.55μm. In addition, the length L_bend of the curved waveguide portion including the length of 300μm of SOA16 is 500μm, and the value of the angle θwg is 5°. Therefore, the waveguide offset ΔY is 44μm, and the waveguide distance D is 88μm. The two AXEL chips (two semiconductor chips 31) are separated by a cleavage process. As mentioned above, the cleavage process usually produces a position offset error of about ±10μm. Similar to Example 1, in this embodiment, the positional offset error in the cleavage process is also taken into consideration, and the narrow waveguide L_nar is designed to be 30 μm, ensuring that the cleavage error compensation area L_scr is greater than 20 μm. In addition, in this embodiment, the chip width (length in the Y direction) is reduced to 200 μm, which is difficult to achieve in the previous AXEL. This is because the constraints on chip size caused by waveguide offset are alleviated by the present disclosure as described above. With respect to the AXEL chip in this embodiment, the straight waveguide portion is arranged at the center of the Y direction in the chip, and margins 1 and margins 2 of 100 μm are respectively ensured on both sides of the Y direction of the straight waveguide portion. The width of previous AXEL chips was limited to about 300 μm, but according to the invention of this embodiment, the number of AXEL chips that can be produced on the same wafer has been successfully increased by 50% by reducing the chip size.

[0094] Next, the basic characteristics of the manufactured AXEL chip were evaluated. At an operating temperature of 55°C, 80mA and 100mA currents were applied to the DFB laser 11 and SOA 16, respectively. In the EA modulator 12, a modulation based on a 10Gbit / s NRZ signal was performed with an applied voltage of -1.5V and an amplitude voltage Vpp=1.5V. It was confirmed that the maximum optical output as modulation reached a high output characteristic of 11dBm. When evaluating the manufacturing yield of the module equipped with the AXEL chip of this embodiment, it was confirmed that the manufacturing yield of the module in the AXEL chip of this embodiment with a narrow waveguide area was about 60%, which was a very good result, relative to the target value of the optical output during modulation of 10dBm or more.

[0095] (Example 3)

[0096] Reference Fig.19 , 20And 21 illustrate Example 3. In this embodiment, in order to improve the manufacturing stability of the window structure portion, a structure that expands the width of the waveguide near the end is adopted. First, the problem related to the manufacturing stability to be solved by this embodiment is explained. In the AXEL of this embodiment, the passive waveguide in the waveguide 32 terminates inside the chip. In the manufacturing process of AXEL, the collapse and disappearance of the passive waveguide will occur starting from the end. The collapse and disappearance of the passive waveguide is one of the reasons that affect the manufacturing yield of AXEL. In addition, the problems caused by the end of the passive waveguide occur more significantly in a narrow waveguide such as the one used in the present disclosure. Refer to Fig.19 This reason will be used for explanation.

[0097] Fig.19 (a) is a cross-sectional view showing the shape of a mesa formed by dry etching in the manufacturing process of AXEL. Fig.19 The cross section of (a) is a diagram of the mesa structure observed from the Y' direction intersecting the Y direction at an angle θwg. Fig.19 (a) and Fig. 9 (b) is the same figure. In the actual manufacturing process of AXEL, a wet etching process is performed to remove the damaged layer of the side wall of the mesa shape generated during dry etching before the embedded regrowth, so that the side wall is exposed to the normal semiconductor surface.

[0098] Fig.19 (b) is a cross-sectional view showing the shape of the mesa after wet etching. Fig.19 The same as (a), Fig.19 The cross section of (b) is a diagram of the mesa structure observed from the Y' direction intersecting the Y direction at an angle θwg. In this wet etching process, a plurality of etching solutions are usually used for the process, but one of them is generally treated with a solution that selectively etches the core layer. As a result, not only the passive waveguide, but also the core layer of the DFB laser 11, the core layer of the SOA 16 (MQW1), and the core layer of the EA modulator 12 (MQW2) are selectively etched at the same time. As a result, in particular in the light emitting region of the DFB laser 11, the damaged layer on the side wall of the mesa structure is removed, thereby suppressing non-luminescent recoupling that does not contribute to oscillation during current injection, improving luminous efficiency, and enhancing light output characteristics and long-term reliability. As Fig.19 As shown in (b), in the passive waveguide, the waveguide core 83a is also selectively etched by the wet etching process. As a result, etching progresses from both sides of the waveguide core 83a, and the core width (length in the Y' direction) becomes narrower relative to the width W1 of the upper part of the mesa structure. In the case of using a waveguide narrower than the usual width as described in the present disclosure, the width of the remaining waveguide core 83a after the wet etching process is further narrowed. Therefore, waveguide anomalies are prone to occur in the AXEL manufacturing process. Here, Fig.19 (c) is a diagram showing the end section of the waveguide near the end face. Fig.19 (c) is a diagram of the mesa structure observed from the direction of light propagation (X' direction orthogonal to Y'). The aforementioned side etching also progresses at the end of the waveguide. Therefore, near the end of the waveguide, since the effect of the etching solution progresses from the two sides of the waveguide and the three directions of the end surface, the side etching is easy to carry out, and the position of the waveguide core 83a becomes a state of being greatly retreated in the direction of SOA16 relative to the original waveguide end position. In particular, this phenomenon is significantly manifested in the case of using a waveguide width narrower than usual as disclosed in the present invention. Starting from the waveguide end that is retreated from the end position of the waveguide core 83a, the mesa structure waveguide is damaged and disappears. The damage and disappearance of the mesa structure waveguide leads to a reduction in the manufacturing yield of the AXEL chip. In order to fundamentally prevent the damage and disappearance of the waveguide, a method of shortening the wet etching processing time before embedding regrowth is considered. However, the wet etching process is a necessary step to ensure the high output and high reliability of the aforementioned DFB laser unit, and therefore it is difficult to solve the problem simply by adjusting the manufacturing conditions.

[0099] Therefore, in this embodiment, a structure is adopted in which the narrow width of the narrow waveguide at the cleavage position 34 of the AXEL chip is converted to a wide width again in the tapered region provided near the end of the waveguide. This structure is a structure that can suppress the breakage and disappearance of the waveguide caused by side etching during the wet etching process, thereby stabilizing the manufacturing process.

[0100] Fig. 20 : This is a diagram showing the schematic structure of the AXEL chip of the present embodiment. In the present embodiment, the yield of the AXEL chip having broadband characteristics capable of high-speed modulation of 50 Gbaud level and high output characteristics of more than 10 dBm can be improved. However, the improvement of the manufacturing yield as an effect of the present embodiment does not depend on the operating speed and light output characteristics of the AXEL chip, and can be introduced into AXEL chips of various specifications. In order to obtain a sufficient reflection suppression effect as described above, the width W2 of the waveguide 32 is narrowed near the cleavage position 34. Thus, the same reflection suppression effect as the aforementioned embodiment 1 can be expected. For the narrowing of the width of the waveguide 32, a tapered region is used to gradually widen the width of the waveguide along the waveguide direction of the light. The AXEL chip 31 of the present embodiment constituting the optical transmitter is a semiconductor chip in which a DFB laser 11 with a length of 300 μm, an EA modulator 12 with a length of 125 μm, and an SOA with a length of 150 μm are monolithically integrated. Furthermore, the waveguide 32 adopts an embedded heterostructure using semi-insulating InP that can achieve high heat dissipation effect and current confinement effect.

[0101] In addition, in this embodiment, in order to cope with the high-speed modulation operation of 50Gbaud, the quantum well structure of the EA modulator 12 and the length of the EA modulator 12 are optimized. As described above, in order to obtain a sufficient reflection suppression effect, the AXEL chip of this embodiment is provided with a tapered region 32c and a narrow waveguide region 32d in the curved waveguide portion 32b. The narrow waveguide region 32d is located near the exit end face. The AXEL chip is formed on the surface of the InP substrate (100), and is configured with a DFB laser 11 in a manner to output light in the direction of the substrate orientation

[011] . After the light from the DFB laser 11 passes through the EA modulator 12 on the same optical axis as the DFB laser 11, the transmission direction is changed through the curved waveguide portion so that it has an angle θwg relative to the direction of the crystal orientation

[011] , and is incident on the SOA16. The light emitted from SOA16 maintains the angle θwg, is converted to the width W2 of the narrow waveguide in the tapered region, passes through the predetermined cleavage position 44, and is widened along the transmission direction of the light through another tapered region arranged in the adjacent chip, and is converted to a wide waveguide of W3, and then reaches the waveguide end of the extended waveguide portion 32g. Here, as a bending angle that can obtain a sufficient reflection suppression effect, the angle θwg is set to 5°. In addition, in order to obtain a sufficient reflection suppression effect, the width W2 of the narrow waveguide region is set to 0.5μm as a range that can be stably manufactured in semiconductor processing technology.

[0102] Under this condition, the narrow waveguide region does not reach the cutoff condition. Therefore, light is transmitted in the narrow waveguide region without radiating. Moreover, the two AXEL chips adjacent to each other in the X direction are opposite to each other. In addition, in the two AXEL chips, the straight waveguide parts in which the DFB laser 11 and the EA modulator 12 are formed are arranged in parallel and with the optical axis aligned. The curved waveguide parts of the two AXEL chips are arranged in parallel and are arranged at a separation distance D at the predetermined cleavage position. Here, the length L_bend of the curved waveguide part of the AXEL chip is 400μm. Therefore, the waveguide distance D is 70μm. The two AXEL chips are separated by the cleavage process. As mentioned above, the cleavage process usually produces a positional offset error of about ±10μm. Therefore, the light emission position of the final chip is determined at the moment when the cleavage process is completed.

[0103] It is necessary to ensure that the length of the narrow waveguide region after the tapered region in the AXEL chip is 5 μm ( Fig.13 In addition, considering the positional deviation error in the cleavage process, it is necessary to ensure that the cleavage error compensation area L_scr is greater than 20μm ( Fig.13 M1), the length of the residual waveguide in the adjacent AXEL chip is guaranteed to be more than 5μm ( Fig.13M3). Therefore, in this embodiment, the length L_nar of the narrow waveguide is designed to be 30μm. When the position offset of the cleavage process is set to 0, a 15μm narrow waveguide portion remains in the AXEL chip. In addition, even if the cleavage position offset occurs and the narrow waveguide area becomes shorter, as long as the cleavage error is within the range of ±10μm, a narrow waveguide area of ​​more than 5μm will remain in the chip. On the contrary, even if the narrow waveguide area becomes longer due to the cleavage position offset, the length of the narrow waveguide area in the chip is less than 25μm, and the length of the residual waveguide remaining in the oppositely configured AXEL chip must be more than 5μm.

[0104] In addition, the extended waveguide portion 32g of the waveguide 32 is located at a position far enough away from the predetermined cleavage position so that the light transmitted in the AXEL chip is not affected. As for the size of the extended waveguide portion 32g, the length L_wide is set to 5μm and the width W3 is set to 4μm. For L_wide and W3, light will not be transmitted, so there is no need to consider optical characteristics during design. In short, from the perspective of the manufacturing stability of the AXEL chip, the side etching caused by the wet etching process is not easy to progress, and the mesa structure that collapses and disappears is not easy to occur during the processing process. In addition, the change in the structure of the waveguide end of this embodiment can be achieved by changing the insulating layer mask 85 ( Fig.19 ) is carried out, and has the characteristic of requiring almost no additional production steps and the introduction load of increased cost.

[0105] In this trial production, in order to compare the effect, the waveguide end width was narrow (W3 = W2 = 0.5 μm) and wide (W3 = 4 μm) in the same batch and process, and the comparison was made. First, the side etching amount after wet etching was compared in the mesa processing process during the production process. In the normal production process, the side etching amount after wet etching was compared in the waveguide cross-sectional direction ( Fig.19 (b)), side etching progresses to about 0.25μm on one side. Therefore, in order to achieve a waveguide width W2 = 0.5μm, the mesa width during the dry etching process for mesa formation needs to be pre-set to 1.0μm. In the case of a waveguide with a narrow width at the end of the waveguide, when the amount of side etching in the cross section relative to the direction of travel of the light at the end is confirmed, it is 0.6μm. Since etching progresses from three directions at the end of the waveguide, the side etching will go very deep and is likely to cause the collapse and disappearance of the waveguide core. In contrast, when the width W3 of the waveguide is expanded to 4μm, the amount of side etching in the cross section relative to the direction of travel of the light at the end is reduced to 0.3μm. By expanding the width W3 of the waveguide, in addition to reducing the amount of side etching at the end of the waveguide, the area of ​​the remaining waveguide core 83a is also increased, thus becoming a physically stable structure. Therefore, it is not easy to cause damage to the waveguide starting from the end of the waveguide.

[0106] Fig.21 The graph shows the results of investigating the relationship between the width W3 of the waveguide end and the side etching amount in the samples produced on the same wafer. In this embodiment, the width W3 of the waveguide end is set to 4 μm. Fig.21 It was confirmed that the amount of side etching can be sufficiently suppressed when the width W3 of the waveguide terminal portion is 3 μm or more.

[0107] Next, the actual AXEL manufacturing process was completed, and the appearance inspection of the AXEL chips made from the same wafer was performed to compare the chip yields. Here, a normal appearance inspection was performed to compare the proportion of faulty chips caused by waveguide damage starting from the waveguide end. As a result, in the AXEL chip of this embodiment (the width W3 of the waveguide end = 4μm), it was confirmed that the yield of faulty chips caused by waveguide damage was improved by about 15% compared with the AXEL chip of Example 1 (the width W3 of the waveguide end = W2 = 0.5μm).

[0108] Next, the basic characteristics of the manufactured AXEL chip were evaluated. At an operating temperature of 55°C, 60mA and 100mA currents were applied to the DFB laser 11 and SOA 16, respectively. In the EA modulator 12, a 50Gbit / s NRZ signal was modulated with an applied voltage of -1.5V and an amplitude voltage Vpp=1.5V. It was confirmed that the maximum output of the light during modulation was 10.5dBm. Even when the width of the waveguide end was expanded, no effect on the basic characteristics of the AXEL chip was found, and the performance was confirmed to be the same as before.

[0109] (Example 4)

[0110] Reference Fig. 22 Example 4 is described below. In this example, the amount of side etching is suppressed by adjusting the crystal plane orientation of the waveguide end, thereby improving manufacturing stability. As described above, in the AXEL chip disclosed herein, the passive waveguide ends inside the chip, and the collapse and disappearance of the waveguide generated from the end portion during the processing step becomes a cause of reduced yield. The cause of the collapse and disappearance of the waveguide is that due to the reference Fig.19 In the wet etching process described in (b), etching progresses on the side surfaces of the waveguide core 83a, and particularly, the etching progress is remarkable at the end portion of the waveguide.

[0111] Therefore, in this embodiment, in addition to expanding the width of the waveguide end shown in the aforementioned embodiment 3, the surface of the waveguide end is set to a (011) crystal plane. The basic structure of the AXEL chip is the same as that of the aforementioned embodiment 3, and in this embodiment, the surface of the waveguide end is changed.

[0112] Fig. 22 1 is a top view showing a schematic structure near the end of the waveguide in the AXEL chip of this embodiment. In order to obtain a sufficient reflection suppression effect, a narrow waveguide region 32d with a narrow width W2 is provided in the waveguide 32 near the cleavage position. Thus, the same reflection suppression effect as in the aforementioned embodiment 1 or 3 can be expected. A tapered region is also provided in the waveguide 32 so that the width of the waveguide gradually changes to a narrow width W2 along the direction of light travel. The AXEL chip is formed on the surface of an InP substrate (100), and a DFB laser 11 is configured to output light in the direction of substrate orientation

[011] . After the light from the DFB laser 11 passes through the EA modulator 12 on the same optical axis as the DFB laser 11, the transmission direction is changed by the curved waveguide portion so that it has an angle θwg relative to the direction of the crystal orientation

[011] , and is incident on the SOA 16. The light emitted from SOA16 maintains an angle θwg, is converted to a narrow waveguide width W2 in the tapered region, passes through the predetermined cleavage position 44, and is widened along the light transmission direction through another tapered region configured in the adjacent chip, and is converted to a wide waveguide of W3 before reaching the waveguide end of the extended waveguide portion 32g. Here, the end face of the extended waveguide portion 32g of the waveguide 32 is a side wall surface formed by dry etching, and the waveguide end face is usually formed as a surface perpendicular to the optical axis. Therefore, when the waveguide itself is at an angle of θwg relative to the crystal orientation

[011] , the end face becomes a surface inclined by θwg from the crystal plane (011).

[0113] In this embodiment, the end face of the extended waveguide portion 32g is not a face perpendicular to the optical axis, but a face that is consistent with the crystal orientation (011) plane. The end face of the extended waveguide portion 32g can be changed by changing the insulating layer mask 85 ( Fig.19 ), and has the characteristic of requiring almost no additional manufacturing process and the introduction load of increased cost. The amount of side etching of the waveguide core 83a based on wet etching described above depends on the surface orientation. That is, the amount of side etching at the (011) plane is the smallest, and there is a tendency that the amount of side etching increases as the angle relative to the (011) plane increases. As described above, at the end of the waveguide, since the side etching progresses from three directions, the side etching progresses most significantly. In this embodiment, by setting the end face of the extended waveguide portion 32g corresponding to the end face of the waveguide to the (011) plane, the side etching of the waveguide core 83a can be suppressed, the collapse and disappearance of the waveguide in the middle of the manufacturing process can be prevented, and stable manufacturing and improved yield can be achieved.

[0114] The AXEL chip of this embodiment was produced and its basic characteristics were evaluated. The structure and manufacturing process of the AXEL chip were the same as those of the AXEL chip shown in Example 3, except for the structure of the waveguide end. In addition, the basic characteristics of the produced AXEL chip also confirmed the same degree of high output and modulation characteristics as those of the AXEL chip of Example 3. In addition, it was confirmed that the chip defect rate caused by abnormalities in the waveguide end portion of the AXEL chip of this embodiment was reduced by about 5% compared with the AXEL chip of Example 3, which was sufficiently effective in improving the manufacturing yield.

[0115] Industrial Applicability

[0116] The manufacturing stability and reflection suppression effect of the optical transmitter can be improved.

[0117] Description of Reference Numerals

[0118] 10: EADFB laser; 11: DFB laser; 12: EA modulator; 13: active layer; 14: diffraction grating; 15: absorption layer; 16: SOA; 20: AXEL; 31: semiconductor chip; 32: waveguide; 32a: straight waveguide portion; 32b: curved waveguide portion; 32c: tapered region; 32d: narrow waveguide region; 32e: waveguide end; 32f: residual waveguide; 32g: extended waveguide portion; 33: exit end face; 34: cleavage position; 35: window region; 44: predetermined cleavage position; 81: InP substrate; 82: lower cladding; 83: core layer; 83a: waveguide core; 84: upper cladding; 85: insulating layer mask; 86: current blocking (InP) layer.

Claims

1. A light emitter having the following components monolithically integrated on a substrate: A distributed feedback DFB laser having an active region formed by a multiple quantum well generating optical gain by current injection and a diffraction grating; An electric field absorption type EA modulator having an absorption region formed of a multi-quantum well having a composition different from that of the DFB laser; A semiconductor amplifier SOA having an active region of the same composition as the DFB laser; Bend the waveguide to rotate the light transmission direction by an angle θwg; as well as A passive waveguide is connected to the SOA and has a core having a bandgap wavelength shorter than the oscillation wavelength of the DFB laser, wherein: The passive waveguide comprises a tapered region and a narrow waveguide region, The tapered region is configured to convert the width W1 of the passive waveguide connected to the SOA into the width W2 of the narrow waveguide region, and W1>W2, The passive waveguide forms an angle θwg with a normal line of the end surface of the substrate and is in contact with the end surface of the substrate.

2. The optical transmitter according to claim 1, further comprising: The second passive waveguide, As for the second passive waveguide, is located at a distance D from the passive waveguide, having a core of the same composition as said passive waveguide, The normal line of the end surface of the substrate is in contact with the end surface of the substrate at an angle θwg, terminated within said substrate, The width is W2, The length is L_d.

3. The optical transmitter according to claim 2, wherein: The substrate is an InP substrate, The DFB laser, the EA modulator and the SOA are formed on the (100) surface of the InP substrate. The optical axis of the DFB laser is in the direction of the substrate crystal orientation [011], The passive waveguide forms an angle θw with the substrate crystal orientation [011] and is connected to the end face of the substrate.

4. The optical transmitter according to claim 3, wherein: As for the second passive waveguide, The width at the position where it contacts the end surface of the substrate is W2, There is an extended waveguide at the end, The width of the extended waveguide portion is W3, W2<W3.

5. The optical transmitter according to claim 4, wherein: W3>3μm.

6. The optical transmitter according to claim 5, wherein: The end face of the extended waveguide portion is consistent with the crystal plane (011) of the substrate.

7. The optical transmitter according to claim 2, wherein: 5μm<L_d<25μm.

8. The optical transmitter according to claim 1, wherein: 0.4μm<W2<0.8μm.

9. The optical transmitter according to claim 1, wherein: 4°<θwg<8°.

10. The optical transmitter according to claim 2, wherein: 30μm<D<140μm.

Images